Download Neuronal polarity: from extracellular signals to

REVIEWS
Neuronal polarity: from extracellular
signals to intracellular mechanisms
Nariko Arimura and Kozo Kaibuchi
Abstract | After they are born and differentiate, neurons break their previous symmetry,
dramatically change their shape, and establish two structurally and functionally distinct
compartments — axons and dendrites — within one cell. How do neurons develop their
morphologically and molecularly distinct compartments? Recent studies have implicated
several signalling pathways evoked by extracellular signals as having essential roles in a
number of aspects of neuronal polarization.
Dendritic spine
A small (sub-micrometre)
membranous extrusion that
protrudes from a dendrite and
forms one half of a synapse.
Filopodia
Finger-like exploratory cell
extensions found in crawling
cells and growth cones. They
are composed largely of
filamentous actin bundles.
Department of Cell
Pharmacology, Graduate
School of Medicine,
Nagoya University, 65,
Tsurumai, Showa, Nagoya,
Aichi 466-8550, Japan.
Correspondence to K.K.
e-mail: kaibuchi@med.
nagoya-u.ac.jp
doi:10.1038/nrn2056
The axons and dendrites of neuronal cells differ from
each other in the composition of their proteins and
organelles. Axons are typically long and thin, with a
uniform width, and they branch at right angles from the
cell body. Dendrites are relatively short; as they emerge
from the cell body they appear thick, but become thinner with increased distance from the cell body and then
undergo Y-shaped branching1,2. Axons contain synaptic
vesicles from which they release neurotransmitters at axon
terminals in response to electrical signals from the cell
body. Dendrites, especially dendritic spines, contain receptors for these neurotransmitters, as well as organelles and
signalling systems3. These two distinct cellular structures
are fundamental for neuronal function, as they enable
neurons to receive and transmit electrical signals.
However, the molecular mechanisms that underlie this
neuronal polarization were unclear until a decade ago.
Many experiments using cultured embryonic hippocampal neurons have revealed that, as they develop,
neurons initially generate several equivalent neurites, but
then begin to polarize so that one neurite becomes an axon
while the remaining neurites become dendrites4 (FIG. 1; see
below). This early asymmetric neurite outgrowth (axon
specification) is regulated by signalling molecules that
have established roles in cytoskeletal rearrangements and
protein trafficking. A balance of positive and negative signals regulates these cellular functions, which result in the
formation of a single axon. Here we provide an overview
of recent progress in the study of axon specification.
Development of neuronal polarization
Cultured rat embryonic hippocampal neurons have
frequently been used to study the establishment of
neuronal polarity. Banker and colleagues developed a
culture system of embryonic hippocampal neurons and
described in detail the morphological changes that occur
194 | MARCH 2007 | VOLUME 8
during polarization4. They divided the morphological
events into five stages (FIG. 1). First, after being dissociated from embryonic rat brains, hippocampal neurons
form several thin filopodia (stage 1). After several hours,
the neurons form a number of immature neurites,
so-called ‘minor processes’ (stage 2). These neurites are
morphologically equal, and undergo repeated, random
growth and retraction. Half a day after plating, one of
these minor processes begins to extend rapidly, becoming much longer than the other neurites (stage 3; see
Supplementary information S1 (movie))5,6. This extended
process becomes an axon; the other minor processes
continue to undergo brief spurts of growth and retraction, maintaining their net length, for up to a week, when
they then become mature dendrites (stage 4). During this
process, dendrites become thicker and shorter than the
axon and begin to establish dendritic components and to
construct premature dendritic spines (stage 5). When the
axon and dendrites are mature, neurons form synaptic
contacts that enable the transmission of electrical activity. Although culturing hippocampal neurons has became
the most popular tool to monitor neuronal polarity, the
morphological changes seen in immature neurons in different brain areas, tissues and organisms are not necessarily
identical in vivo and in vitro (BOX 1).
Specification by intrinsic signalling
How is a single axon specified among equally potential
neurites in cultured hippocampal neurons? A possible
scenario is that signalling factors (morphogens) form
positive and negative feedback loops that can interchangeably affect one another7. Neurite extension and
retraction are controlled by positive and negative signalling molecules respectively, which diffuse within neurites
to regulate the behaviour of actin filaments and microtubules, or to influence cargo transport7. The extension of
www.nature.com/reviews/neuro
© 2007 Nature Publishing Group
REVIEWS
a
Stage 1
Stage 2
Stage 3
Stage 4
Dendrite
Immature
neurite
Stage 5
Dendritic
spine
Axon
Growth
cone
b
c
d
Stage 1
Stage 3
Stage 5
Figure 1 | Dramatic change in neuronal morphology during neuronal
polarization. a | Schematic representation of neuronal polarization in cultured rat
embryonic hippocampal neurons. Shortly after plating, the neurons form small
protrusion veils and a few spikes (stage 1). These truncated protrusions have growth
cones at their tips, and develop into several immature neurites (stage 2). One neurite
then starts to break the initial morphological symmetry, growing at a rapid rate, and
immediately establishing the polarity (stage 3). A few days later, the remaining neurites
elongate and acquire the characteristics of dendrites (stage 4). Approximately seven
days after plating, neurons form synaptic contacts through dendritic spines and axon
terminals, and establish a neuronal network (stage 5). b–d | Immunocytochemical
images of the neurons in stages 1, 3 and 5. The arrowheads indicate the axon. Red
staining represents actin filaments, whereas microtubules are shown by green staining.
The blue staining in (d) shows synapsin 1, which concentrates at the presynaptic
terminal. Panel a modified, with permission, from REF. 4 © (1988) Society for
Neuroscience.
neurites is driven by four main steps: an increase in the
amount of plasma membrane (by vesicle recruitment and
fusion); an increase in the local concentration and activation of signalling molecules (such as phosphatidylinositol
3-kinase (PI3K) and Rho GTPase) and their receptors;
an increase in the dynamics of actin filaments; and the
enhancement of microtubule formation (FIG. 2).
Immediately after a small amount of growth, the
opposite reaction is induced: microtubule catastrophe, a
decrease in the dynamics of actin filaments and enhanced
endocytosis occur. The study of neurite motility indicates that these signalling cascades do indeed form an
extension–retraction cycle in the minor processes (FIG. 2).
Moreover, each neurite seems to release negative feedback signals that feedback to the cell body or metabolize
molecules that are required for axon specification, either
of which might antagonize axon specification in other
neurites7.
Before polarization, therefore, positive and negative
signals seem to be intricately balanced. When this balance is broken by a positive cue, this leads to the activation
of a continuous self-activation system (positive feedback
loop)7,8, and a single minor process elongates to become
an axon. Concurrently, this self-activation system creates strong negative feedback signals that prevent other
neurites from forming a second axon7. Direct evidence
for the existence of negative feedback signals comes from
the observation that growth and/or elongation of a single
neurite driven by laminin-coated beads is inhibited by
the contact with — and subsequent elongation of — a
second neurite with laminin beads (see below)9. This
result indicates the potential existence of negative feedback signals that propagate from the tips of neurites to the
cell body; such signals could be induced by the activation
of transmembrane receptors. This hypothesis addresses
the issue of why only a single neurite is selected to become
an axon, and indicates that promoting neurite elongation
might lead to neuronal polarization.
Box 1 | Similarities and differences among different cell types, and between in vivo and in vitro
Rho GTPase
This enzyme has a molecular
weight of ~21 kDa and
generally serves as a molecular
switch for various cellular
signalling events.
Advances in live imaging techniques have revealed that just after cell division, neurons recognize the asymmetric
extracellular milieu and move towards their destined position with a leading process at the front side and a tailing
process at the rear side. The timing of axon and dendrite formation varies greatly in each type of neuron. For
example, the Purkinje cells in the cerebellum, like the pyramidal cells in the neocortex or hippocampus, are
generated near the ventricular zone, grow an axon (but not minor processes) towards the basal surface, and migrate
toward pre-determined positions following the leading processes that enwrap the glial fibres139. Young cerebellar
granule cells migrate along the pia matter, and then form axons bilaterally and migrate towards the inner layer.
In both Purkinje cells and cerebellar granule cells, nascent dendrites develop at a later stage139. In C. elegans, the
hermaphrodite-specific motorneuron bears several short protrusions, and finally establishes a single axon but not
multiple dendritic processes15. Given the variety of polarized morphologies and the different modes of neuronal
migration in every organism, it is clear that cell-type-specific components play a crucial part. Therefore, at present,
it is difficult to explain every type of neuronal polarization by components and pathways identified from studies of
hippocampal cultures.
More importantly, a recent study compared the morphological change in the early development of retinal ganglion
cells in vivo and in vitro, and showed that in vivo, axons emerge directly from polarized cells in the absence of other
immature neurites that are observed in vitro125. It is possible that this discrepancy of morphological changes between
in vivo and in vitro is caused by the presence of the extracellular signals; these signals could inhibit immature neurite
formation (FIG. 2). The common denominator of polarization in vivo and in vitro seems to be the breaking of symmetry,
which first allows the compartmentalization of the nascent axon, and finally leads to the molecularly and
morphologically distinct axon and dendrite both in vitro and in vivo2. Further analyses both in vivo and in vitro are
required for a better understanding of the mechanism leading to neuronal polarization.
NATURE REVIEWS | NEUROSCIENCE
VOLUME 8 | MARCH 2007 | 195
© 2007 Nature Publishing Group
REVIEWS
a Stage 2
Negative regulation
• Membrane elimination
• Degradation of proteins
• Decrease in dynamics of F-actin
• Microtubule catastrophe
Retraction
Phosphatase
Rho GAP
Negative
feedback signals
b Stage 3
Negative
feedback signals
Axon
specification
Rho GTPases and GEF
PI3K
Centrosome
Extension
Positive regulation
• Membrane recruitment
• Protein transport
• Increase in dynamics of F-actin
• Microtubule assembly
Positive
feedback
loop
Extracellular signals,
receptors,
adhesion molecules.
Transport of
key regulators
Figure 2 | A tentative model for axon specification in neuronal polarization.
a | During stage 2, immature neurites extend and retract randomly to maintain their
overall length. Neurite extension is driven by four main steps: an increase in the amount
of plasma membrane by vesicle recruitment and fusion; the local concentration and
activation of signalling molecules (such as Rho GTPase and phosphatidylinositol 3-kinase
(PI3K)), an increase in actin dynamics; and an increase in microtubule formation. After
extension, some signalling molecules (such as GTPase-activating proteins and
phosphatases) counteract the positive regulation, induce microtubule catastrophe,
decrease actin dynamics and decrease the amount of plasma membrane by endocytosis
and by preventing vesicle fusion. b | When the balance between positive and negative
signals is upset (in the transition from stage 2 to stage 3) by extracellular signals, (auto-)
activation of receptors or adhesion molecules and by the recruitment of signalling
molecules, one neurite elongates rapidly. Continuous elongation is supported by a
positive feedback loop and sustains the activation cycle. The inhibitory signals that
mutually antagonize neurite extension (negative feedback signals; red arrows in a and b)
are progressively generated at the growing axon more than at the other neurites (a thicker
red arrow in b), and interfere with their specification into axons. F-actin, filamentous actin;
GAP, GTPase-activating protein; GEF, guanine nucleotide exchange factor.
Extracellular matrix
Any material that is part of a
tissue yet not part of any cell.
The main components of the
extracellular matrix are various
glycoproteins, proteoglycans
and hyaluronic acid.
Cell-autonomous
A function not driven by
extracellular signals, but only
by an intrinsic program within
cells.
Planar polarization
The coordinated organization
of cells within the plane of a
single-layered sheet of cells.
Plus end
The site of microtubule
polymerization. It allows the
assembly of GTP-bound tubulin
dimers.
Flexibility in specification. Even after the axon is specified, other neurites have the potential to change their
fate during development. Goslin et al. reported that
transecting the axons of hippocampal neurons early
in development could cause the polarity to alternate;
a minor process starts to become an axon, instead of
the transected axon8. This change depends on the difference in the length of the cut axon compared to the
neurites. If the cut axon is at least 10 μm shorter than
the other minor processes, the shortened axon becomes
a dendrite, and the longest minor process begins to grow
rapidly and eventually becomes an axon8. Axon–dendrite
plasticity is known to be retained in stage 4 neurons10,
indicating that, to some degree, the identities of axons
and dendrites are flexible during development.
Extracellular signals and polarity
Neurons in the brain and nervous system are surrounded
not only by other cells, but also by extracellular matrix, and
these extracellular signals are potential cues for specifying
axon formation. It has been reported that UNC-6 (also
known as netrin in mammals), a secreted protein that
mediates axon guidance and cell migration11–14, functions in
196 | MARCH 2007 | VOLUME 8
axon formation to orient asymmetric neuronal growth
in Caenorhabditis elegans15. In wild-type C. elegans, the
hermaphrodite-specific motorneuron (HSN) extends
multiple ventrally directed neurites from its cell body
during the third larval stage (L3), and eventually directs
a single axon ventrally and then anteriorly. In unc-6 or
unc-40 (a receptor for UNC-6) mutants, neurons abnormally extend several immature neurites in all directions during L3, and finally form a single axon directed
anteriorly. Interestingly, another repulsive guidance
cue, semaphorin 3A, is also known to act as a dendritic
chemoattractant and regulates the oriented growth of
apical dendrites in cortical neurons16. Axon guidance cues
seem to stimulate and orient the asymmetrical growth of
neurites and guide the mature axon.
Another secreted protein, WNT, has been reported to
determine neuronal polarity along the anterior–posterior
body axis17,18. WNT and its receptor frizzled (FZ) control
various developmental processes, including asymmetric
cell division, cell fate determination and tissue polarity19.
In C. elegans, the mechanosensory posterior lateral microtubule neuron (PLM) extends a long anterior process that
forms a chemical synapse, and a shorter posterior process
that does not form a synapse. In lin-44/wnt or lin-17/fz
mutants, the longer PLM process extends posteriorly to its
cell body instead of anteriorly. The synaptic vesicle protein
synaptobrevin 1 (SNB-1) in PLM is abnormally enriched
in the longer posterior process in lin-44 or lin-17 mutants.
Interestingly, when the posterior and anterior processes
are identical in length in lin-17 mutants, synaptic vesicles are
localized in both PLM processes17, indicating that these
PLM neurons have two functional axon-like processes.
Given that the lin-44/wnt defect causes LIN-17/FZ protein to mislocalize cell-autonomously18, secreted LIN-44/
WNT seems to act directly on the PLM mechanosensory
neuron through LIN-17/FZ localization, and to specify
the anterior process as an axon.
WNT signalling cascades inhibit the activity of
glycogen synthase kinase 3β (GSK3β) during planar
polarization19,20. In migrating mammalian astrocytes, cell
division cycle 42 (Cdc42)-dependent inactivation of
GSK3β induces the interaction of adenomatous polyposis
coli (APC) protein with microtubule plus ends specifically at the leading edges, and makes the centrosome turn
towards the front side of migrating cells in a dynein- or
dynactin-dependent manner21. It might be interesting to
examine whether extracellular signals such as WNT also
regulate neuronal polarity through GSK3β and dynein.
The neurotrophic factors nerve growth factor
(NGF), brain-derived neurotrophic factor (BDNF) and
neurotrophin 3 (NT3) are also candidate extracellular
polarity-regulating cues, and have been proposed to
accelerate neuronal polarization by enhancing axon
growth. PI3K is known to mediate downstream signalling
from tyrosine receptor kinases (Trks), which are receptors
of neurotrophic factors22. It has been reported that the
ganglioside-converting enzyme plasma membrane
ganglioside sialidase (PMGS), which promotes the activation of TrkA at the tips of neurites, is necessary for axon
outgrowth23. As PMGS accumulates in a single immature
neurite during stage 2, its activity seems essential for the
www.nature.com/reviews/neuro
© 2007 Nature Publishing Group
REVIEWS
Centrosome
The main microtubule
organizing centre, which
consists of two centrioles
surrounded by pericentriolar
material. It is responsible for
microtubule nucleation and
anchoring, and forms the radial
microtubule organization.
Pleckstrin homology (PH)
domain
A sequence of ~100 amino
acids that is present in many
signalling molecules and that
binds to lipid products of phosphatidylinositol 3-kinase.
Ubiquitin–proteasome
system
Ubiquitin is a small regulatory
protein. Ubiquitylation refers
to the post-translational
modification of a protein by
the covalent attachment of one
or more ubiquitin monomers.
After a chain of at least four
ubiquitin peptides are attached
to a lysine on a substrate
protein, the substrates are
degraded by the proteasome.
polarized activation of TrkA and the subsequent activation of PI3K. How PMGS becomes restricted to the single
process before polarization is still unclear; nevertheless,
these results indicate that receptors of neurotrophic
factors are crucial in neuronal polarization.
Many extracellular matrix and cell adhesion molecules, such as laminin or neuron–glia cell adhesion
molecule (NgCAM), enhance neurite outgrowth and
axon elongation. One study tested whether these growthpromoting proteins could selectively enhance axon elongation and specification24. When neurons were plated on
alternating stripes of poly-d-lysine (which promotes cell
adhesion through ionic interactions) and either laminin
or NgCAM, any minor processes growing on the stripe
coated with laminin or NgCAM elongated rapidly and
became an axon, whereas the processes growing only
poly-d-lysine remained shorter. Similar substrate preference was also observed in an experiment using laminincoated beads in stage 2 neurons9. Extracellular signals
might therefore accelerate the growth of minor processes
to determine the neurite identity (BOX 1).
Intracellular signalling and polarity
The observations described above indicate that extracellular cues lead to a polarized intracellular response
that allows neurons to determine axon and dendrite fate during development. However, embryonic
neurons in culture (in the absence of a gradient of physiological extracellular cues) also polarize in a few days.
Therefore, cell-autonomous signalling cascades, which
are also stimulated in response to extracellular signals,
are thought to underlie neuronal polarization.
PI3K and phosphatidylinositol-3,4,5-trisphosphate.
Accumulating evidence indicates the importance of
PI3K and its lipid product, phosphatidylinositol-3,4,5trisphosphate (PtdIns(3,4,5)P3), in neuronal polarization9,15,23,25–27 (BOX 2; TABLE 1). PI3K and PtdIns(3,4,5)P3
exhibit a polarized activation and distribution during
chemotaxis in neutrophils28,29, Dictyostelium discoideum30,31 and fibroblasts32, and this polarity is required
for protrusion of the leading edge membrane in cell
migration33,34. In cultured hippocampal neurons, PI3K
Box 2 | Key regulators and experimental criteria in neuronal polarity
Among the various molecules involved in neuronal polarity, which molecules have a
central role? Accumulating evidence indicates that evolutionarily conserved proteins
such as small GTPases, phosphatidylinositol 3-kinase (PI3K) and the partitioning defect
(PAR) complex control neuronal polarity. We list in TABLE 1 the prominent molecules
reported as regulators of neuronal polarity, with their subcellular distribution and the
effect of up- or downregulated activity. The criteria used to score the effect on
polarization varies greatly. For the integration of information, we suggest that it would
be better to develop some common criteria. The axon or dendrite formation should be
determined by several specific markers, such as tau 1, synaptic markers and cytoskeletal
markers (TABLE 1). In particular, morphology should not be used as the only determinant
for axons, and a minor neurite should not be called a ‘dendrite’ before differentiation.
Early neuronal polarization is divided into 3 steps: symmetry breaking and elongation
of a selective neurite; stabilization of the broken symmetry by a positive or negative
feedback loop; and compartmentalization of the axon and somatodendrite. It will be of
great importance to examine which molecules are coupled to each step.
NATURE REVIEWS | NEUROSCIENCE
activity concentrates at the tip of the newly specified
axon during stage 3 (REF. 25), and inhibiting PI3K activity prevents axon specification9,25,26. Local activation
of PI3K generates PtdIns(3,4,5)P3 at the membrane,
which recruits proteins that contain the complementary
pleckstrin homology (PH) domain. The local activation of
minor processes using laminin-coated beads results
in the accumulation of PtdIns(3,4,5)P3 at the tips of
immature neurites, and induces the dramatic growth
of an axon from a neurite9. Overexpression of Pten
(phosphatase and tensin homologue deleted on chromosome 10), the product of which dephosphorylates
PtdIns(3,4,5)P3, disrupts the development of polarity25.
These results indicate that the polarized activation of
PI3K and the accumulation of PtdIns(3,4,5)P3 at the tip
of a single minor process (future axon) are required for
axon specification.
PI3K, Akt and GSK3 β . Stimulation of developing
neurons with BDNF or NT3 also activates the PI3K–
PtdIns(3,4,5)P3–Akt–GSK3β signalling pathway (TABLE 1).
In response to stimulation by neurotrophic factors, PI3K
is activated through Ras (FIG. 3) and insulin receptor substrate (IRS)22,35. PI3K, through PtdIns(3,4,5)P3, activates
phosphoinositide-dependent kinase 1 (PDK1)36. PDK1,
and other kinases such as integrin-linked kinase, phosphorylates the protein kinase Akt (also known as protein kinase B) at threonine 308 and serine 473 (REF. 37).
Activated Akt phosphorylates GSK3β on serine 9. GSK3β
is constitutively active, and its phosphorylation by Akt
inactivates the kinase. As Akt localizes to the tips of the
axons in hippocampal neurons, the phosphorylated, inactive GSK3β is also restricted at the tips of growing axons
in cultured hippocampal neurons38,39. Overexpression of
constitutively active GSK3β disrupts axon elongation,
whereas knockdown of GSK3β and the use of specific
inhibitors or transfection of the constitutively active form
of Akt into neurons38,40 cause the formation of multiple
axons38,39, indicating that the activity of Akt and GSK3β
regulates neuronal polarity. One study has reported that
the inhibition of GSK3β at later stages of development
causes dendrites to change suddenly to axons38. This indicates the significance of the PI3K–PtdIns(3,4,5)P3–Akt–
GSK3β signalling pathway in neuronal polarity, especially
during the transition from minor processes and dendrites
to axons. However, a recent study used neurons derived
from knock-in mice in which the two isoforms of GSK3
(GSK3α and GSK3β) were unable to be phosphorylated
by Akt to show that GSK3β inhibition during neuronal
polarization is not mediated by Akt-dependent phosphorylation41. Although the authors have not mentioned
which molecules are responsible for GSK3β inhibition,
membrane transport and sorting machinery might be
involved in GSK3β-dependent neuronal polarization41
(BOX 2; TABLE 1).
How are these signalling molecules localized at the
tip of the axon? Interestingly, it has been reported that
inhibition of the ubiquitin–proteasome system induces
symmetric Akt distribution and multiple axons42. This
indicates that localized protein degradation regulates
protein stability and neuronal polarization.
VOLUME 8 | MARCH 2007 | 197
© 2007 Nature Publishing Group
REVIEWS
Table 1 | Neuronal polarity-regulating molecules in cultured hippocampal neurons
Protein (function)
Subcellular
localization
PI3K (kinase; produces Activited at the tip of
PtdIns(3,4,5)P3)
future axon (st.2,3)
*Effect of
upregulated activity
Myr-p110: multiple
axons
PTEN (PtdIns(3,4,5)P3
degradating enzyme)
Uniform (EGFP–
PTEN)
WT: unpolarized
R-Ras (small GTPase)
Single neurite (st.2);
axon (st.3)
WT, DA (Q87L):
multiple axons
H-Ras (small GTPase)
ND
RAP1B (small GTPase)
Tip of single minor
process (st.2); tip of
axon (st.3)
Cdc42 (small GTPase)
Tip of single minor
process (st.2); tip of
axon (st.3)
Rac1 (small GTPase)
ND
Akt (mediates the
signals of growth
factors)
Tips of minor
processes (st.2);
tip of axon (st.3)
GSK3β (glycogen
synthesis)
Tips of all neurites
(st.2); tip of axon
(st.3); pGSK3β: tip of
axon (st.3); cell body
PAR3 (involved in
asymmetric cell
division and cell
polarization)
PAR6 (involved in
asymmetric cell
division and cell
polarization)
aPKC (involved in
asymmetric cell
division and cell
polarization)
SAD (presynaptic
differentiation)
Tips of minor
processes (st.2);
tip of axon (st.3)
MARK2 (microtubule
affinity-regulating
kinase)
CRMP2 (mediator
in semaphorin 3A
signalling; cargo
receptor)
*Effect of downregulated activity
Inhibitor (LY294002
etc.): no axon, or
delay of polarization
siRNA: increased
number of axons
siRNA:
unpolarization;
R-RasGAP-Myr:
prevention of axon
formation
WT, DA (V12): multiple DN(N17): no axon
axons
siRNA: loss of
DA (V12): superpolarity, no axon
numerary axons, or
reduced number of
minor processes
siRNA: no axon;
DA (L28):
supernumerary axons; N17: no effect
V12: unpolarized
DA (V12): increased
the length of
minor neurite, or
unpolarized
WT, Myr-Akt: multiple
axons, or increased
number
of neurites
WT, S9A: no axon or
unpolarized
Criteria used to
score effects
Morphological and
molecular (tau-1)
Association with
other molecules
PtdIns(3,4,5)P3, Akt
Morphological and
molecular (PAR3,
tau-1, MAP2, pAkt,
synapsin 1, GAP43)
Morphological and
molecular (tau-1,
synapsin 1, GAP43,
FM4-64 uptake, MAP2,
pAkt, pGSK3β)
Morphological and
molecular (tau-1)
Morphological and
molecular (tau-1,
MAP2, PAR3, p-aPKC,
pAkt)
Morphological and
molecular (tau-1,
MAP2, PAR3, p-aPKC,
pAkt, synapsin 1)
Morphological and
molecular (tau-1,
MAP2, synapsin 1)
PI3K, Akt, GSK3β
Refs
9,25,
38,40
25,38
GSK3β, Akt
51
PI3K, MEK, GSK3β,
CRMP2
PI3K, Cdc42, PAR3,
PAR6, aPKC, Akt
40
26
PI3K, PAR3, PAR6,
aPKC, Akt
26,63
STEF
26,63
Morphological and
molecular (tau-1,
MAP2, pAkt, synapsin
1, GAP43)
Morphological
shRNA, inhibitor
and molecular
GID5–6 peptide:
(tau-1, MAP2, pAkt,
multiple axon
synapsin 1, GAP43,
GSK3α/β knock-in
synaptophysin)
mice: no effect
delC, delN: no axon, Morphological and
or increased length molecular (tau-1,
of minor processes MAP2, synapsin 1)
PI3K, Ras, GSK3β,
tau, MAP1B
38,40,
42
PAR3, PI3K, Akt,
CRMP2, tau, BDNF,
NT3, MAP1B, APC
38,39,
41,42,
83
PAR6, aPKC, STEF
25,26,
63
WT, delCREB:
unpolarized
Morphological and
molecular (tau-1,
MAP2, synapsin 1)
PAR3, aPKC,
Cdc42, STEF
25,26,
63
25,26,
63
DN (N17): reduced
the length of
minor process, or
unpolarized
shRNA: neuronal
death
Tip of axon (st.3)
WT, 4N-1:
unpolarized, or
multiple axon-like
neurites
ND
ND
ND
Inhibitor (Bis):
unpolarized
Morphological and
molecular (tau-1,
synapsin 1, MAP2)
PAR3, PAR6
Diffuse (st.3)
ND
Morphological and
molecular (tau-1,
MAP2)
Tau
94
All neurites, or tip of
longest neurite (st.2);
tip of axon (st.3)
Diffuse (st.2); distal
part of axon (st.3)
WT or kinase dead
mutant: unpolarized
KO: immature
axon formation,
enhancement of
dendritic formation
siRNA: multiple
axon-like neurites
Tau, aPKC
89
WT: multiple axons,
or increased length
of axon
Morphological and
molecular (tau-1,
synapsin 1)
siRNA: unpolarized, Morphological
or decreased length and molecular
(tau-1, synapsin 1,
of axon
synaptophysin, MAP2)
Tubulin, kinesin 1,
numb, GSK3β, Akt,
PI3K, BDNF, NT3
39,98,
99,104,
132
*Descriptions of ‘effect of up- or downregulated activity’ were extracted from each reference. 4N-1, deletion construct containing first coiled–coil domain of PAR3;
APC, adenomatosis polyposis coli; BDNF, brain-derived neurotrophic factor; Bis, bisindolylmaielmide; CRMP2, collapsin response mediator protein; DA, dominant active;
delCREB, CRIB domain deletion mutant; delC, C-terminal deletion mutant; delN, N-terminal deletion mutant; DN, dominant negative; EGFP, enhanced green fluorescent
protein; FM4-64, fluorescent membrane marker; GAP43, growth-associated protein 43; KO, knockout mice; L28, mutant where amino acid at 28 has been replaced with
Leu; MAP1B, mitogen-activated protein 1B; MARK2, microtubule affinity-regulating kinase 2; MEK, mitogen-activated protein kinase kinase; Myr-p110, myristylated p110,
which is the subunit of PI3K containing the kinase domain; N17, mutant where amino acid at 17 has been replaced with Asn; ND, not determined; NT3, neurotrophin 3;
pAkt, phosphorylated Akt; p-aPKC, phosphorylated atypical protein kinase C; PAR3, partitioning defect 3; pGSK3β, phosphorylated glycogen synthase kinase 3β; PI3K,
phosphatidylinositol 3-kinase; PtdIns(3,4,5)P3, phosphatidylinositol-3,4,5-trisphosphate; PTEN, phosphatase and tensin homologue deleted on chromosome 10; Q87L,
amino acid Gln in position 87 is mutated to Leu; RAP1B, Ras-related protein 1b; S9A, constitutive active mutant with Ser in position 9 mutated to Ala; shRNA, short hairpin
RNA; siRNA, small interfering RNA; st., stage; STEF, SIF and TIAM1-like exchange factor; V12, mutant where amino acid at 12 has been replaced with Val; WT, wild type.
198 | MARCH 2007 | VOLUME 8
www.nature.com/reviews/neuro
© 2007 Nature Publishing Group
REVIEWS
RNA interference
(RNAi). A method by which
double-stranded RNA that is
encoded on an exogenous
vector can be used to interfere
with normal RNA processing,
causing rapid degradation of
the endogenous RNA and
thereby precluding translation.
This provides a simple way of
studying the effects of the
absence of a gene product in
simple organisms and in cells.
Guanine nucleotide
exchange factor
(GEF). A protein that facilitates
the exchange of GDP for GTP in
the nucleotide-binding pocket
of a GTP-binding protein.
Adaptor protein
A protein that contributes to
cellular function by recruiting
other proteins to a complex.
Such molecules often contain
several protein–protein
interaction domains.
PI3K and RAP1. Ras-related protein 1B (RAP1B), a Ras
superfamily GTPase, localizes to the tips of prospective
axons before the accumulation of Cdc42 and the partitioning defect (PAR) complex (TABLE 1; see below)26.
Rap1b overexpression induces multiple axon-like
neurites and the accumulation of the PAR complex in
each neurite. Suppression of Rap1b expression by RNA
interference (RNAi) causes the complete loss of axons.
The concentration of RAP1B into a single neurite
seems to contribute to the recruitment of essential molecules for axon specification, and therefore could be a
decisive step.
The Rap1b RNAi-induced axonal loss is partially
rescued by expressing an active form of Cdc42, whereas
axonal loss in response to treatment with a PI3K inhibitor is rescued by the active form of RAP1B. So, RAP1B
seems to function upstream of Cdc42 (REF. 43) and the
PAR complex, and downstream of PI3K, probably in a
PtdIns(3,4,5)P3-dependent manner44 (FIG. 3). It is noteworthy that RAP1 is activated by the neurotrophin receptor TrkA, which is itself activated at the tip of growing
axons23. Active TrkA recruits guanine nucleotide exchange
factors (GEFs) of RAP1 (such as C3G) and adaptor
proteins (such as CRK and fibroblast growth factor
receptor substrate 2), and stimulates RAP1 (REFS 45–49).
Further studies are needed to elucidate how RAP1B is
concentrated and activated in prospective axons.
Small GTPases of the Ras family. Ras proteins (H-Ras,
K-Ras, N-Ras and R-Ras), which are small GTPases
that regulate cell growth and differentiation50, are also
reported as activators of PI3K40,51 during neuronal polarization. Active, GTP-bound Ras interacts with several
effector proteins; the best characterized are PI3K and
Raf 52–55 (FIG. 3). In D. discoideum, active Ras and PI3K
localized at the leading edge regulate directional cell
movement, indicating that Ras regulates cell polarity56.
Consistent with this, an essential role for Ras in neuronal
polarization has been reported40,51. Overexpression of
wild-type Ras induced multiple axons in cultured hippocampal neurons, whereas ectopically expressed dominantnegative Ras inhibited axon formation40,51. Inhibition of
PI3K or mitogen-activated protein kinase kinase (MEK),
which lies downstream of Raf, suppresses the Ras-induced
formation of multiple axons. This indicates that Ras
functions upstream of PI3K and the mitogen-activated
protein kinase (MAPK) cascade in the establishment and
maintenance of neuronal polarity.
Dominant-negative
A mutant form of a molecule
that can form a heteromeric
complex with the normal
molecule, knocking out the
activity of the entire complex.
Growth cones
Hand-like structures at the tip
of growing neurites, axons,
immature neurites and
dendrites.
Cytochalasin D
An actin depolymerizing
chemical compound.
Rho GTPases and their regulators
During initiation of axon growth, the growth cone
undergoes net elongation and retraction driven by the
rapid polymerization and depolymerization of actin
filaments57–59. Applying cytochalasin D to stage 2 neurons
causes multiple axon-like neurites59, indicating that the
reorganization of actin filaments is necessary for determining a single axon. Among many other molecules,
Rho GTPases have been reported as important regulators of axon specification26,60–63. Activated Rho GTPases
activate specific effector molecules to change cell
morphology or motility64.
NATURE REVIEWS | NEUROSCIENCE
Among the 18 types of Rho GTPase, RhoA, Cdc42
and Rac1 have been the most extensively characterized65.
In neuroblastoma cells, such as PC12 cells or N1E-115
cells, the activation of Cdc42 or Rac1 enhances neurite
elongation, whereas RhoA activation is associated with
inhibition of neurite formation66–69. NGF-induced neurite outgrowth in PC12 cells involves several Rho GTPase
effector molecules: p21-activated kinase (PAK)70, p35 (a
neuron-specific regulator for cyclin-dependent kinase 5)71,
myotonic dystrophy kinase-related Cdc42-binding kinase
(MRCK)72, neural Wiskott–Aldrich syndrome protein
(N-WASP)73, and IQGAP3 (REF. 74). These effector
molecules influence neurite outgrowth through the
rearrangements of microtubules or actin filaments.
Rho GTPase regulators. Numerous Rho GTPase
regulatory mechanisms are implicated in neuronal
polarization. The Rac1-specific GEF T-lymphoma and
metastasis 1 protein (TIAM1)75 and SIF and TIAM1like exchange factor (STEF; also known as TIAM2)76
have been identified as regulators of neuronal polarity
upstream of Rac1 (REFS 62,63) (FIG. 3). In hippocampal
neurons, TIAM1 accumulates at the tip of the prospective axon. Overexpression of Tiam1 induces multiple
axon-like neurites, whereas the depletion of Tiam1
inhibits axon formation by preventing actin filament
reorganization62. Interestingly, a distinct pathway has
been delineated in which local activation of a novel
Rac-specific GEF, dedicator of cytokinesis 7 (DOCK7),
regulates Rac activity to inactivate the microtubuledestabilizing protein stathmin (also known as OP18)
and promote axon formation77. The DOCK7 protein
is asymmetrically distributed in non-polarized stage 2
hippocampal neurons and selectively expressed in the
axon. Knockdown of Dock7 expression prevents axon
formation, whereas overexpression induces the formation of multiple axons77. So, DOCK7, and TIAM1 or
STEF, trigger the activation of Rac to regulate microtubule and actin remodelling, respectively. In addition,
it has been reported that the cytoplasmic dynein light
chain TCTEX1 functions in initial neurite sprouting
and axon outgrowth through the activation of Rac61. So,
Rac and its GEFs are necessary for establishing neuronal
polarity in response to upstream signals.
The PAR complex regulates polarity
The prototypic par genes were identified in C. elegans
for their roles in directing asymmetric cell division
during early development78. The PAR protein complex,
PAR3–PAR6–aPKC (atypical protein kinase C), is crucial
for establishing and maintaining the anterior–posterior
polarity at the single-cell embryo stage in C. elegans,
as well as in epithelial cells and neuroblast cells of
Drosophila melanogaster. In hippocampal neurons, the
PAR3–PAR6–aPKC complex becomes concentrated
into a single axon from all minor processes during the
transition between stages 2 and 3 (REFS 25,79) (TABLE 1).
Inhibition of aPKC activity prevents axon formation25,
whereas phosphorylated (active) aPKC can be seen at the
tips of growing axons26. Furthermore, inhibitors of PI3K
prevent polarization and cause mislocalization of PAR3
VOLUME 8 | MARCH 2007 | 199
© 2007 Nature Publishing Group
REVIEWS
Guidance cues
WNT
Netrin
Receptors
Adhesion molecules
Plasma membrane
Positive signal
Negative signal
Rho GEF
Ras
PI3K
RhoA
ILK
PAK
Rac1
RAP1B
Rho-kinase
Raf
Positive feedback
loop
PIP3
PTEN
PDK1
IQGAP3
GAP
Cdc42 GEF
Akt
N-WASP
Cdc42
p35/CDK5
MAPK
Rac GEF
PAR6
aPKC
PAR3
GSK3β
MEK
MRCK
SRA1
(STEF/TIAM1)
WAVE 1
MAPKAP-K1
MARK2
LIMK
APC
CRMP-2
Tau
MAP1B
Microtubules
KLC
Transport
β-Catenin
MLC
Cell adhesion
Arp2/3
Cofilin
F-actin
Stathmin
CREB
L1
Gene expression
Figure 3 | Simplified model of positive and negative signals. In response to the activation of cell surface receptors by
extracellular ligands or adhesion molecules, Ras becomes activated. Ras activates phosphatidylinositol 3-kinase (PI3K) and
other signalling molecules. PI3K-mediated production of phosphatidylinositol-3,4,5-trisphosphate (PtdIns(3,4,5)P3; or PIP3)
tethers PtdIns(3,4,5)P3-binding molecules and activates Ras-related protein 1b (RAP1B) and integrin-linked kinase (ILK), or
phosphoinositide-dependent kinase 1 (PDK1). RAP1B activates cell division cycle 42 (Cdc42) at the tips of growing axons.
As partitioning defect 6 (PAR6) specifically associates with the GTP-bound form of Cdc42, the PAR3–PAR6–aPKC (atypical
protein kinase C) complex becomes activated. PAR3 can then interact with, and activate, the Rac guanine nucleotide
exchange factors (GEFs) T-lymphoma and metastasis 1 protein (TIAM1) and SIF and TIAM1-like exchange factor (STEF), and
thereby activate Rac1. Active Rac1 and Cdc42 interact with effector molecules, such as neural Wiskott–Aldrich syndrome
protein (N-WASP), myotonic dystrophy kinase-related Cdc42-binding kinase (MRCK), p21-activated kinase (PAK), IQGAP3,
p35–cyclin-dependent kinase 5 (CDK5), and specifically Rac1-associated protein 1 (SRA1)1– WASP-family verprolinhomologous protein 1 (WAVE1), and regulate filamentous actin (F-actin) reorganization. As Rac1 can activate PI3K, this
forms a positive feedback loop. Specific GTPase-activating proteins (GAPs) for Rac1 and Cdc42 will inactivate both
GTPases respectively and therefore can break this feedback loop. Activated ILK or PDK1 phosphorylates Akt (also known
as protein kinase B) which, in turn, phosphorylates and inactivates glycogen synthase kinase 3β (GSK3β). Microtubule
affinity-regulating kinase 2 (MARK2) is inhibited by aPKC. As GSK3β or MARK2 activity promotes microtubule
destabilization, inactivation of either of these kinases has a positive effect on microtubule formation and neurite
outgrowth. Inactive GSK3β also activates phosphatase and tensin homologue deleted on chromosome 10 (PTEN). In some
cases, Rho is activated in response to certain ligands or guidance cues including WNT and netrin, and targets PTEN to
membranous areas and enhances its activity. Active PTEN decreases the levels of PtdIns(3,4,5)P3 at the leading edge of the
neurite, and prevents positive regulatory signal transduction. The function of some downstream molecules of Rho kinase,
GSK3β and mitogen-activated protein kinase-activated protein kinase 1 (MAPKAP-K1) — such as β-catenin, myosin light
chain (MLC) and cyclic AMP-response element binding protein (CREB) — were examined in non-neuronal cells, and their
exact function in neuronal polarity needs to be addressed in the future. APC, adenomatosis polyposis coli; ARP2/3, actinrelated protein 2/3; CRMP2, collapsin response mediator protein 2; KLC, kinesin light chain; LIMK, LIM domain kinase;
MAP1B, microtubule-associated protein 1B; MEK, mitogen-activated protein kinase kinase.
200 | MARCH 2007 | VOLUME 8
www.nature.com/reviews/neuro
© 2007 Nature Publishing Group
REVIEWS
APP-containing
vesicles
WAVE 1
APP
SRA1
GAP43
JIPs
PIP3-containing
vesicles
Tubulin
hetrodimer
α β
CRMP2
aPKC
CRMP2
PIP3
PAR6
PAR3
PIP3BP
KAP3
KLC
KIF5
KIF3
–
GAKIN
+
Rac1 by activating STEF. Ectopic expression of Stef and
Par3 induces multiple axon-like neurites but impairs
axon maturation63, indicating that axonally localized
Rac1 and PAR3 activity is required for axon formation
but is not sufficient for axon maturation. Given that
GTP-bound Rac1 can activate PI3K84–86, this might
form the basis for the positive feedback loop that results
in the continued activation of PI3K and downstream
molecules (FIG. 3).
Microtubule
Centrosome
Figure 4 | Transport of key regulators in axon formation by kinesin. When one
neurite begins to adopt the fate of an axon, kinesin family member 5 (KIF5) strictly
concentrates in the growing axon. Consistent with this, several key molecules that are
involved in axon formation start to accumulate into the distal parts or tips of axons. These
molecules are directly or indirectly associated with kinesins. Cargo, such as amyloid
precursor protein (APP), growth-associated protein 43 (GAP43), the SRA1–WAVE1
complex (where SRA1 is specifically Rac1-associated protein 1 and WAVE1 is WASPfamily verprolin-homologous protein 1) and tubulin heterodimers, that cannot directly
interact with kinesins, binds to ‘cargo receptors’ such as c-Jun amino-terminal kinase
(JNK)-interacting proteins (JIPs) and collapsin response mediator protein 2 (CRMP2),
which subsequently dock to kinesin light chain (KLC). The molecules are transported
specifically towards the axon terminal and promote axon specification. The partitioning
defect (PAR3)–PAR6–aPKC (atypical protein kinase C) complex is transported towards
the axon terminal by kinesin 2 (which comprises KIF3 and kinesin superfamily-associated
protein 3 (KAP3)), through the direct binding of PAR3 to KAP3. Phosphatidylinositol3,4,5-trisphosphate (PtdIns(3,4,5)P3) is associated with PtdIns(3,4,5)P3-binding protein
(PIP3BP), and is transported by the guanylate kinase-associated kinesin (GAKIN) to the
prospective axon.
and PAR6 (REF. 25), indicating that the localization and
activity of the PAR complex are regulated downstream
of PI3K, and are required for neuronal polarization. A
study in D. melanogaster indicates that PAR6, aPKC and
bazooka (as PAR3 is known in D. melanogaster) are not
involved in axon or dendrite specification in this species80. Physiological circumstance or another, unknown
signal cascade in D. melanogaster might compensate
for the PAR complex-induced phenotype observed in
cultured rat hippocampal neurons (BOX 2).
Lamellipodia
Sheet-like extensions at the
edge of a cell that contain a
crosslinked filamentous actin
meshwork, and that are often
associated with cell migration.
Microtubule-associated
proteins
(MAPs). MAPs, including tau
and MAP1B, are high
molecular weight proteins that
comprise four distinct
polypeptide chains. MAPs are
distributed along the lengths of
microtubules, and have a
stabilizing role.
From Cdc42 to Rac through PAR. Recent data indicate
that the PAR complex mediates Cdc42-induced Rac activation through STEF or TIAM1 (REF. 63) in hippocampal
neurons (FIG. 3). The PAR complex associates with GTPbound Cdc42 more than with GTP–Rac1 through a
Cdc42/Rac-interactive binding (CRIB) domain of
PAR6 (REFS 81,82) (FIG. 3; TABLE 1). Recent research has
shown that STEF and TIAM1 directly interact with the
carboxy (C)-terminal region of PAR3 (REFS 63,83) and
form a complex with PAR3–PAR6–aPKC–GTP-bound
Cdc42 (REF. 63). The isolated STEF-binding fragment of
PAR3 is sufficient to induce lamellipodia independently
of Cdc42 and PAR6, indicating that PAR3 can activate
NATURE REVIEWS | NEUROSCIENCE
PAR complex localization. The PAR complex is transported from the cell body to the axon terminal by the
plus-end-directed kinesin motor protein kinesin 2
(which comprises kinesin family member 3 (KIF3) and
kinesin superfamily-associated protein 3 (KAP3))79,87,
and begins to concentrate in a growing axon during the
transition from stage 2 to 3 (FIG. 4). Signals, which could
be evoked by upstream molecules (which might also be
transported by motor proteins or the centrosome; see
below), stabilize the PAR complex at the tip of the future
axon, thereby generating a localized positive feedback
loop. This positive feedback loop might be a driving
force for initial axon sprouting and axonal maturation.
By contrast, in epithelial cells and mature neurons, an
opposite role of Rac1 has been reported. In these cells,
Rac1 is activated by the depletion of PAR3 by RNAi83,88.
One possible explanation is that knockdown of PAR3 in
these cells seems to prevent the recruitment of TIAM1
to the cell–cell contact site between epithelial cells or to
the dendritic spine in neurons, and thereby promotes the
inappropriately localized activation of Rac1.
PAR complex and MARK2. Another function of the PAR
complex in neuronal polarization has been reported
recently89 (FIG. 3). Microtubule affinity-regulating kinase
2 (MARK2; also known as PAR1b) is a serine/threonine
protein kinase that phosphorylates microtubule-associated
proteins (MAPs) such as tau90. The depletion of MARK2
by RNAi was reported to induce multiple axons in
hippocampal neurons89 (TABLE 1). By contrast, ectopic
expression of MARK2 increases tau phosphorylation
and prevents axon formation. This defect is rescued by
the ectopic expression of the PAR complex, as aPKC
phosphorylates MARK2 on threonine 595 and inactivates its kinase activity91–93. A point mutant of MARK2,
in which the aPKC phosphorylation site is replaced by
alanine, also causes the defect of axon formation, but this
defect cannot be corrected by the ectopic expression of
PAR3, PAR6 or aPKC. These results indicate that aPKC
in complex with PAR3 and PAR6 negatively regulates
MARK2 activity at the tip of the axon to prevent MAP
phosphorylation and thereby assemble microtubules.
A study of SAD kinases — which phosphorylate tau
and contain a kinase domain related to that of MARK2
(REF 94) — found that the neurons from SAD kinaseknockout mice showed a loss of axons and abnormally
orientated dendrites. SAD kinase is neither phosphorylated nor downregulated by aPKC89. Although their
mechanism of regulation is distinct, both MARK2 and
SAD kinase seem to regulate neuronal polarity through
microtubule reorganization.
VOLUME 8 | MARCH 2007 | 201
© 2007 Nature Publishing Group
REVIEWS
CRMP2 and axon elongation
The collapsin response mediator protein (CRMP; also
known as TOAD, ULIP and DRP) family (FIG. 3,4), which
displays significant homology to C. elegans UNC-33, is
expressed at high levels in the developing nervous system95,96. unc-33 mutants exhibit severely uncoordinated
movement and abnormalities in axon guidance and axon
growth in many neurons97. We have shown that CRMP2
is involved in the regulation of axon formation during
neuronal polarization98. CRMP2 is enriched in the growing axons of cultured hippocampal neurons during stage
3, and overexpression of Crmp2 induces the formation of
multiple axons. As the induced axon-like processes possess synaptophysin-positive synaptic terminals, CRMP2
is thought to induce and to maintain the mature surplus
axon98. Furthermore, overexpression of Crmp2 can alter
an established dendrite to become an axon during stage 4,
indicating that overexpressed Crmp2 confers axonal identity not only on immature neurites but also on established
dendrites98. These observations indicate that CRMP2 has
a crucial role in axon formation of hippocampal neurons,
thereby establishing and maintaining neuronal polarity.
Potential CRMP2 effectors. Several CRMP2 binding
proteins have been identified. CRMP2 interacts with
tubulin heterodimers and promotes microtubule assembly in vitro99. It binds and colocalizes with numb proteins
— which function in Notch-induced cell fate determination100,101 and in clathrin-mediated endocytosis102,103
— in central regions of axonal growth cones and, in turn,
regulates the endocytosis of L1, a neuronal cell adhesion
molecule104. CRMP2 associates with the SRA1–WAVE1
complex105 (where SRA1 is specifically Rac1-associated
protein 1 (REF. 106); also known as CYFIP1 (REF. 107)) and
WAVE1 is WASP-family verprolin-homologous protein 1
(REF. 108)), and regulates actin filament stability during
lamellipodia formation105 (FIG. 4). So, CRMP2 seems to
promote neurite elongation and axon specification by
regulating microtubule assembly, endocytosis of adhesion
molecules and reorganization of actin filaments.
GSK3β phosphorylates CRMP2. We and others have
found that GSK3β phosphorylates CRMP2 at threonine
514 and inactivates it39,109,110. In cultured hippocampal
neurons, approximately 30% of CRMP2 is constitutively
phosphorylated at threonine 514, and this phosphorylation is decreased by GSK3β inhibitors. GSK3βphosphorylated CRMP2 is localized in the distal part of
the growing axon, but clearly excluded from the axonal
growth cones39, indicating that there is a pool of nonphosphorylated (on threonine 514) CRMP2 in the growing axonal tips. Phosphorylation of CRMP2 lowers its
ability to interact with tubulin dimer and numb99,111, and
therefore this CRMP2 loses its activity to enhance microtubule formation and L1 endocytosis. The expression of a
non-phosphorylated form of CRMP2 efficiently induces
the formation of multiple axons, and can also counteract
the inhibitory effects of constitutively active GSK3β on
neuronal polarization, indicating that GSK3β regulates
neuronal polarity by phosphorylating CRMP2 (REF. 39).
NT3 and BDNF inhibit GSK3β activity, and, therefore,
202 | MARCH 2007 | VOLUME 8
CRMP2 phosphorylation. CRMP2 depletion prevents
NT3-induced axon outgrowth39. Taken together, the
regulation of GSK3β and CRMP2 activities is of crucial
importance in neuronal polarization.
GSK3β and other MAPs. GSK3β also phosphorylates
tau112,113, MAP1B114 and APC115. These three substrates
stabilize microtubules in the absence of phosphorylation, but are prevented from doing so by GSK3βmediated phosphorylation115,116. One study reported that
APC phosphorylation by GSK3β is prevented downstream
of NGF and PI3K at the tip of the axon in neurons117.
Hippocampal neurons derived from mice lacking both
tau and MAP1B show axon loss at stage 3 (REF. 118). As
microtubule stabilization and protrusion into the actin
meshwork at the distal area of the growth cone promote
axon elongation57,59,119, GSK3β seems to have a central
role as a negative regulator of neuronal polarization, and
inactivation of GSK3β downstream of neurotrophic factors increases the stability of microtubules by increasing
the activity of MAPs and vesicle trafficking41 (FIG. 3). It has
been reported that partial elimination of GSK3α/β activity
results in axon branch formation, whereas complete elimination of GSK3α/β results in a marked inhibition of axon
growth120. GSK3α/β activity might be tightly regulated
spatially and temporally, so that neurites can undergo
dynamic growth and retraction through microtubule
reorganization during early development. In addition, it
has been found that casein kinase 2 and GSK3β synergistically phosphorylate PTEN at its C terminus, which
inhibits its activity121. As GSK3β is constitutively active,
the PTEN activity is probably usually downregulated,
and therefore upregulation of the PTEN activity by
inactivation of GSK3β might provide negative feedback
regulation, which counteracts the PI3K activity before and
after axon specification in minor processes (FIG. 3). Further
work is needed to elucidate the relationship between
GSK3β and PTEN in neuronal polarization.
Microtubules and neuronal polarity
As mentioned earlier, MAPs and regulators of microtubule organization are crucial in neuronal polarization.
How do microtubules regulate axon specification?
The position of centrosomes in the cell body has been
reported to be important for axon determination by
influencing where the nascent axon will eventually protrude122–124. Lefcort and Bentley123 first demonstrated a
correlation between centrosome and axon formation in
grasshopper neurons in 1989, and this was generalized by
using granule cells and hippocampal neurons124. Zmuda
and Rivas122 reported that the centrosome consistently
appeared to be repositioned to the base of the newly
emerging axon during transition from a unipolar to bipolar morphology in granule cells. Another recent paper
shows that neurons that contain multiple centrosomes
have multiple axons; neurons without centrosomes fail
to grow stable axons124. Directional microtubule assembly
and membrane transport towards specific neurites can be
observed in stage 2 neurons124. Therefore, the position of
the centrosome promotes the outgrowth of one neurite to
become an axon over other minor processes.
www.nature.com/reviews/neuro
© 2007 Nature Publishing Group
REVIEWS
Initial segment
The proximal part of an axon,
the neck of the axonal shaft.
It is reported that the position of the centrosome is
regulated by Cdc42, cytoplasmic dynein, GSK3β and
certain microtubule-associated molecules in a variety
of cells21. As some essential cascades for polarity seem
to be conserved from worm to mammal, the position
of the centrosome might be regulated by extracellular
signals and downstream signalling cascades that involve
the PAR complex and GSK3β in developmental neurons.
This issue, however, still remains somewhat ambiguous,
because different studies using distinct cell types and
experimental conditions have reported that the centrosome does not localize to the site of the emerging
axon125 (BOX 1). Further studies are required to more
precisely determine the specific role of the centrosome
in neuronal polarity.
Transport of key regulators
A distinct set of proteins that are responsible for either
axon or dendrite formation contributes to the functional differences in each compartment2,126. Kinesin 1
(previously called ‘kinesin’ or ‘conventional kinesin’) is
a microtubule plus-end-directed motor127,128 that comprises two kinesin heavy chains (KIF5; also known as
KHC) and two kinesin light chains (KLC)129,130 (FIG. 4).
KIF5 contains the motor domain and KLC contains the
binding domain for the cargo (see below). KIF5 preferentially accumulates in a mature axon, especially in
its initial segment, through the association of the motor
domain with microtubules in the initial segment131. This
preferential localization of KIF5 into a single neurite is
observed before neurons become polarized6. KIF5 was
not seen to accumulate in minor processes undergoing
growth spurts in stage 2 neurons. Dominant-negative
KIF5 or the depletion of KLC by RNAi prevent axon
specification132, which indicates that at least the basic
activity of kinesin 1 is required for axonal growth.
Kinesin cargos. The specific nature of kinesin 1 transport
enables the cargos of kinesin 1 to target a single neurite
(FIG. 4). Several cargo proteins have been reported to be
transported by kinesin 1 during the early stages of neuronal development, such as growth-associated protein 43
(GAP43)133 and amyloid precursor protein (APP)134.
c-Jun amino-terminal kinase (JNK)-interacting proteins
(JIPs)135 and CRMP2 (REFS 105,132,136) function as ‘cargo
receptors’ to link other cargo vesicles or molecules with
kinesin 1. The CRMP2–kinesin 1 complex is evolutionarily
1.
2.
3.
4.
5.
6.
Fukata, Y., Kimura, T. & Kaibuchi, K. Axon specification
in hippocampal neurons. Neurosci. Res. 43, 305–315
(2002).
Craig, A. M. & Banker, G. Neuronal polarity. Annu.
Rev. Neurosci. 17, 267–310 (1994).
Nimchinsky, E. A., Sabatini, B. L. & Svoboda, K.
Structure and function of dendritic spines. Annu. Rev.
Physiol. 64, 313–353 (2002).
Dotti, C. G., Sullivan, C. A. & Banker, G. A. The
establishment of polarity by hippocampal neurons in
culture. J. Neurosci. 8, 1454–1468 (1988).
Ruthel, G. & Hollenbeck, P. J. Growth cones are not
required for initial establishment of polarity or
differential axon branch growth in cultured hippocampal
neurons. J. Neurosci. 20, 2266–2274 (2000).
Jacobson, C., Schnapp, B. & Banker, G. A. A change in
the selective translocation of the Kinesin-1 motor
7.
8.
9.
conserved from worms to mammals132,134, and regulates
the transport of tubulin heterodimers or SRA1–WAVE1
to the distal part of the growing axon to influence the
organization of microtubules and actin filaments105,132.
GSK3β phosphorylates KLC and causes the release of
kinesin 1 from membrane-bound organelles137. It will be
important to determine what kind of signals and posttranslational modifications (such as phosphorylation)
regulate the activity of kinesin motors or cargo receptors,
and ensure that cargo transport is tightly regulated.
It has been reported that PtdIns(3,4,5)P3 could be
transported by the guanylate kinase-associated kinesin
(GAKIN) to the prospective axon, where it regulates
neuronal polarity formation27. Similarly, PtdIns(3,4,5)P3
was seen to be transported to the neurite terminal in
laminin bead-induced neurite elongation (N.A. and K.K.,
unpublished observations). Furthermore, a novel protein
shootin 1 is preferentially transported towards axon terminals, where it is required for the activity of PI3K138. So,
two independent mechanisms maintain PtdIns(3,4,5)P3 at
the tips of axons: the production by PI3K at the axon tip,
and the recruitment of PtdIns(3,4,5)P3 by GAKIN. Taken
together, the kinesin family of proteins has important roles
in neuronal polarity by recruiting the key molecules into a
single neurite to establish neuronal polarization.
Conclusion and perspectives
Many papers concerning neuronal polarity using
hippocampal cultures have been published recently. In
particular, evolutionarily conserved proteins, such as PI3Kassociated molecules and small GTPases, have now been
reported to be involved in neuronal polarity, as observed
in other migrating cells or polarized cells (BOX 2). The
function of some of these proteins, for example PI3K,
guidance cues and signalling molecules, have also been
confirmed in vivo16. However, at present the functions
of other molecules implicated in neuronal polarity have
not been fully evaluated and are mainly based on in vitro
data. In order to confirm gene functions in polarity, the
information derived from in vivo systems are becoming
more important (BOX 1). Nevertheless, neuronal cultures
will remain a useful tool to examine neuronal migration
and polarization. In vivo and in vitro studies — when
based on common criteria — will further complement
each other and will lead to a new and integrated insight
for understanding the molecular machinery required to
establish neuronal polarity.
domain marks the initial specification of the axon.
Neuron 49, 797–804 (2006).
Describes that the motor domain of kinesin 1
accumulates in a single neurite before polarization
of a neuron.
Andersen, S. S. & Bi, G. Q. Axon formation: a
molecular model for the generation of neuronal
polarity. Bioessays 22, 172–179 (2000).
An excellent review that extensively discusses the
positive and negative feedback loops in neuronal
polarization.
Goslin, K. & Banker, G. Experimental observations on
the development of polarity by hippocampal neurons
in culture. J. Cell Biol. 108, 1507–1516 (1989).
Menager, C., Arimura, N., Fukata, Y. & Kaibuchi, K.
PIP3 is involved in neuronal polarization and axon
formation. J. Neurochem. 89, 109–118 (2004).
NATURE REVIEWS | NEUROSCIENCE
10. Bradke, F. & Dotti, C. G. Differentiated neurons retain
the capacity to generate axons from dendrites.
Curr. Biol. 10, 1467–1470 (2000).
11. Wadsworth, W. G., Bhatt, H. & Hedgecock, E. M.
Neuroglia and pioneer neurons express UNC-6 to
provide global and local netrin cues for guiding
migrations in C. elegans. Neuron 16, 35–46 (1996).
12. Serafini, T. et al. The netrins define a family of axon
outgrowth-promoting proteins homologous to
C. elegans UNC-6. Cell 78, 409–424 (1994).
13. Hong, K. et al. A ligand-gated association between
cytoplasmic domains of UNC5 and DCC family
receptors converts netrin-induced growth cone
attraction to repulsion. Cell 97, 927–941 (1999).
14. Hedgecock, E. M., Culotti, J. G. & Hall, D. H. The
unc-5, unc-6, and unc-40 genes guide
circumferential migrations of pioneer axons and
VOLUME 8 | MARCH 2007 | 203
© 2007 Nature Publishing Group
REVIEWS
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
mesodermal cells on the epidermis in C. elegans.
Neuron 4, 61–85 (1990).
Adler, C. E., Fetter, R. D. & Bargmann, C. I. UNC-6/
Netrin induces neuronal asymmetry and defines the
site of axon formation. Nature Neurosci. 9, 511–518
(2006).
Whitford, K. L., Dijkhuizen, P., Polleux, F. & Ghosh, A.
Molecular control of cortical dendrite development.
Annu. Rev. Neurosci. 25, 127–149 (2002).
Prasad, B. C. & Clark, S. G. Wnt signaling establishes
anteroposterior neuronal polarity and requires
retromer in C. elegans. Development 133,
1757–1766 (2006).
Hilliard, M. A. & Bargmann, C. I. Wnt signals and
frizzled activity orient anterior-posterior axon outgrowth
in C. elegans. Dev. Cell 10, 379–390 (2006).
Logan, C. Y. & Nusse, R. The Wnt signaling pathway in
development and disease. Annu. Rev. Cell Dev. Biol.
20, 781–810 (2004).
Montcouquiol, M., Crenshaw, E. B., 3rd & Kelley, M. W.
Noncanonical Wnt signaling and neural polarity.
Annu. Rev. Neurosci. 29, 363–386 (2006).
Etienne-Manneville, S. & Hall, A. Rho GTPases in cell
biology. Nature 420, 629–635 (2002).
Huang, E. J. & Reichardt, L. F. Trk receptors: roles in
neuronal signal transduction. Annu. Rev. Biochem. 72,
609–642 (2003).
Da Silva, J. S., Hasegawa, T., Miyagi, T., Dotti, C. G. &
Abad-Rodriguez, J. Asymmetric membrane
ganglioside sialidase activity specifies axonal fate.
Nature Neurosci. 8, 606–615 (2005).
Esch, T., Lemmon, V. & Banker, G. Local presentation
of substrate molecules directs axon specification by
cultured hippocampal neurons. J. Neurosci. 19,
6417–6426 (1999).
Shi, S. H., Jan, L. Y. & Jan, Y. N. Hippocampal
neuronal polarity specified by spatially localized
mPar3/mPar6 and PI 3-kinase activity. Cell 112,
63–75 (2003).
A key paper that describes the effect of PI3K and
the PAR complex on neuronal polarization.
Schwamborn, J. C. & Puschel, A. W. The sequential
activity of the GTPases Rap1B and Cdc42 determines
neuronal polarity. Nature Neurosci. 7, 923–929
(2004).
Horiguchi, K., Hanada, T., Fukui, Y. & Chishti, A. H.
Transport of PIP3 by GAKIN, a kinesin-3 family
protein, regulates neuronal cell polarity. J. Cell Biol.
174, 425–436 (2006).
Rickert, P., Weiner, O. D., Wang, F., Bourne, H. R. &
Servant, G. Leukocytes navigate by compass: roles of
PI3Kγ and its lipid products. Trends Cell Biol. 10,
466–473 (2000).
Servant, G. et al. Polarization of chemoattractant
receptor signaling during neutrophil chemotaxis.
Science 287, 1037–1040 (2000).
Meili, R. et al. Chemoattractant-mediated transient
activation and membrane localization of Akt/PKB is
required for efficient chemotaxis to cAMP in
Dictyostelium. EMBO J. 18, 2092–2105 (1999).
Jin, T., Zhang, N., Long, Y., Parent, C. A. &
Devreotes, P. N. Localization of the G protein βγ
complex in living cells during chemotaxis. Science
287, 1034–1036 (2000).
Haugh, J. M., Codazzi, F., Teruel, M. & Meyer, T.
Spatial sensing in fibroblasts mediated by 3’
phosphoinositides. J. Cell Biol. 151, 1269–1280
(2000).
Van Haastert, P. J. & Devreotes, P. N. Chemotaxis:
signalling the way forward. Nature Rev. Mol. Cell Biol.
5, 626–634 (2004).
Weiner, O. D. Regulation of cell polarity during
eukaryotic chemotaxis: the chemotactic compass.
Curr. Opin. Cell Biol. 14, 196–202 (2002).
Yamada, M. et al. Insulin receptor substrate (IRS)-1
and IRS-2 are tyrosine-phosphorylated and associated
with phosphatidylinositol 3-kinase in response to brainderived neurotrophic factor in cultured cerebral cortical
neurons. J. Biol. Chem. 272, 30334–30339 (1997).
Alessi, D. R. et al. Characterization of a
3-phosphoinositide-dependent protein kinase which
phosphorylates and activates protein kinase Bα.
Curr. Biol. 7, 261–269 (1997).
Burgering, B. M. & Coffer, P. J. Protein kinase B (c-Akt)
in phosphatidylinositol-3-OH kinase signal
transduction. Nature 376, 599–602 (1995).
Jiang, H., Guo, W., Liang, X. & Rao, Y. Both the
establishment and the maintenance of neuronal
polarity require active mechanisms: critical roles of
GSK-3β and its upstream regulators. Cell 120,
123–135 (2005).
39. Yoshimura, T. et al. GSK-3β regulates phosphorylation
of CRMP-2 and neuronal polarity. Cell 120, 137–149
(2005).
This work, together with that of reference 38,
provides the evidence that the PI3K/Akt/GSK3β
signal cascade has essential roles in neuronal
polarization, especially in the determination of
axon and dendrite fate.
40. Yoshimura, T. et al. Ras regulates neuronal polarity via
the PI3-kinase/Akt/GSK-3β/CRMP-2 pathway.
Biochem. Biophys. Res. Commun. 340, 62–68
(2006).
41. Gartner, A., Huang, X. & Hall, A. Neuronal polarity is
regulated by glycogen synthase kinase-3 (GSK-3β)
independently of Akt/PKB serine phosphorylation.
J. Cell Sci. 119, 3927–3934 (2006).
42. Yan, D., Guo, L. & Wang, Y. Requirement of dendritic
Akt degradation by the ubiquitin-proteasome system
for neuronal polarity. J. Cell Biol. 174, 415–424
(2006).
43. Hogan, C. et al. Rap1 regulates the formation of
E-cadherin-based cell–cell contacts. Mol. Cell Biol. 24,
6690–6700 (2004).
44. Lova, P. et al. A selective role for phosphatidylinositol
3,4,5-trisphosphate in the Gi-dependent activation of
platelet Rap1B. J. Biol. Chem. 278, 131–138 (2003).
45. Kao, S., Jaiswal, R. K., Kolch, W. & Landreth, G. E.
Identification of the mechanisms regulating the
differential activation of the mapk cascade by
epidermal growth factor and nerve growth factor in
PC12 cells. J. Biol. Chem. 276, 18169–18177 (2001).
46. Gotoh, T. et al. Identification of Rap1 as a target for
the Crk SH3 domain-binding guanine nucleotidereleasing factor C3G. Mol. Cell Biol. 15, 6746–6753
(1995).
47. Sasagawa, S., Ozaki, Y., Fujita, K. & Kuroda, S.
Prediction and validation of the distinct dynamics of
transient and sustained ERK activation. Nature Cell
Biol. 7, 365–373 (2005).
48. Vossler, M. R. et al. cAMP activates MAP kinase and
Elk-1 through a B-Raf- and Rap1-dependent pathway.
Cell 89, 73–82 (1997).
49. York, R. D. et al. Rap1 mediates sustained MAP
kinase activation induced by nerve growth factor.
Nature 392, 622–626 (1998).
50. Hancock, J. F. Ras proteins: different signals from
different locations. Nature Rev. Mol. Cell Biol. 4,
373–384 (2003).
51. Oinuma, I., Katoh, H. & Negishi, M. R-Ras controls
axon specification upstream of GSK-3β through
integrin-linked kinase. J. Biol. Chem. 282, 303–318
(2007).
52. Dickson, B., Sprenger, F., Morrison, D. & Hafen, E. Raf
functions downstream of Ras1 in the Sevenless signal
transduction pathway. Nature 360, 600–603 (1992).
53. Rodriguez-Viciana, P. et al. Phosphatidylinositol-3-OH
kinase as a direct target of Ras. Nature 370,
527–532 (1994).
54. Rodriguez-Viciana, P., Warne, P. H., Vanhaesebroeck, B.,
Waterfield, M. D. & Downward, J. Activation of
phosphoinositide 3-kinase by interaction with Ras and
by point mutation. EMBO J. 15, 2442–2451 (1996).
55. Vojtek, A. B., Hollenberg, S. M. & Cooper, J. A.
Mammalian Ras interacts directly with the serine/
threonine kinase Raf. Cell 74, 205–214 (1993).
56. Sasaki, A. T., Chun, C., Takeda, K. & Firtel, R. A.
Localized Ras signaling at the leading edge regulates
PI3K, cell polarity, and directional cell movement.
J. Cell Biol. 167, 505–518 (2004).
57. Dent, E. W. & Gertler, F. B. Cytoskeletal dynamics and
transport in growth cone motility and axon guidance.
Neuron 40, 209–227 (2003).
58. Baas, P. W. & Buster, D. W. Slow axonal transport and
the genesis of neuronal morphology. J. Neurobiol. 58,
3–17 (2004).
59. Bradke, F. & Dotti, C. G. The role of local actin
instability in axon formation. Science 283,
1931–1934 (1999).
60. Bito, H. et al. A critical role for a Rho-associated
kinase, p160ROCK, in determining axon outgrowth in
mammalian CNS neurons. Neuron 26, 431–441
(2000).
61. Chuang, J. Z. et al. The dynein light chain Tctex-1 has
a dynein-independent role in actin remodeling during
neurite outgrowth. Dev. Cell 9, 75–86 (2005).
62. Kunda, P., Paglini, G., Quiroga, S., Kosik, K. &
Caceres, A. Evidence for the involvement of Tiam1 in
axon formation. J. Neurosci. 21, 2361–2372 (2001).
63. Nishimura, T. et al. PAR-6-PAR-3 mediates Cdc42induced Rac activation through the Rac GEFs STEF/
Tiam1. Nature Cell Biol. 7, 270–277 (2005).
204 | MARCH 2007 | VOLUME 8
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
This work, together with that of reference 62,
provides evidence that the activity of Cdc42 and
Rac1 is essential for neuronal polarization and
neuronal maturation.
Ridley, A. J. et al. Cell migration: integrating signals
from front to back. Science 302, 1704–1709 (2003).
Govek, E. E., Newey, S. E. & Van Aelst, L. The role of
the Rho GTPases in neuronal development. Genes Dev.
19, 1–49 (2005).
An excellent review that extensively describes a
variety of signalling cascades that mediate Rho
GTPases and the associating factors.
Sebok, A. et al. Different roles for RhoA during neurite
initiation, elongation, and regeneration in PC12 cells.
J. Neurochem. 73, 949–960 (1999).
Kranenburg, O. et al. Activation of RhoA by
lysophosphatidic acid and Gα12/13 subunits in
neuronal cells: induction of neurite retraction.
Mol. Biol. Cell 10, 1851–1857 (1999).
Kozma, R., Sarner, S., Ahmed, S. & Lim, L. Rho family
GTPases and neuronal growth cone remodelling:
relationship between increased complexity induced by
Cdc42Hs, Rac1, and acetylcholine and collapse
induced by RhoA and lysophosphatidic acid. Mol. Cell
Biol. 17, 1201–1211 (1997).
Tigyi, G. et al. Lysophosphatidic acid-induced neurite
retraction in PC12 cells: control by phosphoinositideCa2+ signaling and Rho. J. Neurochem. 66, 537–548
(1996).
Daniels, R. H., Hall, P. S. & Bokoch, G. M. Membrane
targeting of p21-activated kinase 1 (PAK1) induces
neurite outgrowth from PC12 cells. EMBO J. 17,
754–764 (1998).
Nikolic, M., Chou, M. M., Lu, W., Mayer, B. J. &
Tsai, L. H. The p35/Cdk5 kinase is a neuron-specific
Rac effector that inhibits Pak1 activity. Nature 395,
194–198 (1998).
Chen, X. Q., Tan, I., Leung, T. & Lim, L. The myotonic
dystrophy kinase-related Cdc42-binding kinase is
involved in the regulation of neurite outgrowth in
PC12 cells. J. Biol. Chem. 274, 19901–19905
(1999).
Banzai, Y., Miki, H., Yamaguchi, H. & Takenawa, T.
Essential role of neural Wiskott–Aldrich syndrome
protein in neurite extension in PC12 cells and rat
hippocampal primary culture cells. J. Biol. Chem.
275, 11987–11992 (2000).
Wang, S. et al. IQGAP3, a novel effector of Rac1 and
Cdc42, regulates neurite outgrowth. J. Cell Sci. 23 Jan
2006 (doi:10.1242/jcs.03356).
Habets, G. G. et al. Identification of an invasioninducing gene, Tiam-1, that encodes a protein with
homology to GDP–GTP exchangers for Rho-like
proteins. Cell 77, 537–549 (1994).
Hoshino, M. et al. Identification of the stef gene that
encodes a novel guanine nucleotide exchange factor
specific for Rac1. J. Biol. Chem. 274, 17837–17844
(1999).
Watabe-Uchida, M., John, K. A., Janas, J. A.,
Newey, S. E. & Van Aelst, L. The Rac activator DOCK7
regulates neuronal polarity through local
phosphorylation of stathmin/Op18. Neuron 51,
727–739 (2006).
Cowan, C. R. & Hyman, A. A. Asymmetric cell division
in C. elegans: cortical polarity and spindle positioning.
Annu. Rev. Cell Dev. Biol. 20, 427–453 (2004).
Nishimura, T. et al. Role of the PAR-3-KIF3 complex in
the establishment of neuronal polarity. Nature Cell
Biol. 6, 328–334 (2004).
Rolls, M. M. & Doe, C. Q. Baz, Par-6 and aPKC are
not required for axon or dendrite specification in
Drosophila. Nature Neurosci. 7, 1293–1295
(2004).
Lin, D. et al. A mammalian PAR-3–PAR-6 complex
implicated in Cdc42/Rac1 and aPKC signalling and cell
polarity. Nature Cell Biol. 2, 540–547 (2000).
Johansson, A., Driessens, M. & Aspenstrom, P. The
mammalian homologue of the Caenorhabditis elegans
polarity protein PAR-6 is a binding partner for the Rho
GTPases Cdc42 and Rac1. J. Cell Sci. 113,
3267–3275 (2000).
Chen, X. & Macara, I. G. Par-3 controls tight junction
assembly through the Rac exchange factor Tiam1.
Nature Cell Biol. 7, 262–269 (2005).
Tolias, K. F., Cantley, L. C. & Carpenter, C. L. Rho
family GTPases bind to phosphoinositide kinases.
J. Biol. Chem. 270, 17656–17659 (1995).
Keely, P. J., Westwick, J. K., Whitehead, I. P., Der, C. J.
& Parise, L. V. Cdc42 and Rac1 induce integrinmediated cell motility and invasiveness through PI3K.
Nature 390, 632–636 (1997).
www.nature.com/reviews/neuro
© 2007 Nature Publishing Group
REVIEWS
86. Chan, T. O. et al. Small GTPases and tyrosine kinases
coregulate a molecular switch in the phosphoinositide
3-kinase regulatory subunit. Cancer Cell 1, 181–191
(2002).
87. Shi, S. H., Cheng, T., Jan, L. Y. & Jan, Y. N. APC and
GSK-3β are involved in mPar3 targeting to the nascent
axon and establishment of neuronal polarity. Curr.
Biol. 14, 2025–2032 (2004).
88. Zhang, H. & Macara, I. G. The polarity protein PAR-3
and TIAM1 cooperate in dendritic spine
morphogenesis. Nature Cell Biol. 8, 227–237 (2006).
89. Chen, Y. M. et al. Microtubule affinity-regulating
kinase 2 functions downstream of the PAR-3/PAR-6/
atypical PKC complex in regulating hippocampal
neuronal polarity. Proc. Natl Acad. Sci. USA 103,
8534–8539 (2006).
90. Drewes, G., Ebneth, A., Preuss, U., Mandelkow, E. M.
& Mandelkow, E. MARK, a novel family of protein
kinases that phosphorylate microtubule-associated
proteins and trigger microtubule disruption. Cell 89,
297–308 (1997).
91. Suzuki, A. et al. aPKC acts upstream of PAR-1b in
both the establishment and maintenance of
mammalian epithelial polarity. Curr. Biol. 14,
1425–1435 (2004).
92. Hurov, J. B., Watkins, J. L. & Piwnica-Worms, H.
Atypical PKC phosphorylates PAR-1 kinases to
regulate localization and activity. Curr. Biol. 14,
736–741 (2004).
93. Kusakabe, M. & Nishida, E. The polarity-inducing
kinase Par-1 controls Xenopus gastrulation in
cooperation with 14–3-3 and aPKC. EMBO J. 23,
4190–4201 (2004).
94. Kishi, M., Pan, Y. A., Crump, J. G. & Sanes, J. R.
Mammalian SAD kinases are required for neuronal
polarization. Science 307, 929–932 (2005).
95. Arimura, N., Menager, C., Fukata, Y. & Kaibuchi, K.
Role of CRMP-2 in neuronal polarity. J. Neurobiol. 58,
34–47 (2004).
96. Goshima, Y., Nakamura, F., Strittmatter, P. &
Strittmatter, S. M. Collapsin-induced growth cone
collapse mediated by an intracellular protein related
to UNC-33. Nature 376, 509–514 (1995).
97. Hedgecock, E. M., Culotti, J. G., Thomson, J. N. &
Perkins, L. A. Axonal guidance mutants of
Caenorhabditis elegans identified by filling sensory
neurons with fluorescein dyes. Dev. Biol. 111,
158–170 (1985).
98. Inagaki, N. et al. CRMP-2 induces axons in cultured
hippocampal neurons. Nature Neurosci. 4, 781–782
(2001).
A key paper showing that overexpression of Crmp2
induces multiple axons in cultured neurons. Most
importantly, this paper is one of the early reports
showing that the polarity-regulating molecules
could interchange the fate of a neurite from
dendrite to axon.
99. Fukata, Y. et al. CRMP-2 binds to tubulin heterodimers
to promote microtubule assembly. Nature Cell Biol. 4,
583–591 (2002).
100. Frise, E., Knoblich, J. A., Younger-Shepherd, S.,
Jan, L. Y. & Jan, Y. N. The Drosophila Numb protein
inhibits signaling of the Notch receptor during cell–cell
interaction in sensory organ lineage. Proc. Natl Acad.
Sci. USA 93, 11925–11932 (1996).
101. Spana, E. P. & Doe, C. Q. Numb antagonizes Notch
signaling to specify sibling neuron cell fates. Neuron
17, 21–26 (1996).
102. Berdnik, D., Torok, T., Gonzalez-Gaitan, M. &
Knoblich, J. A. The endocytic protein α-Adaptin is
required for numb-mediated asymmetric cell division
in Drosophila. Dev. Cell 3, 221–231 (2002).
103. Santolini, E. et al. Numb is an endocytic protein.
J. Cell Biol. 151, 1345–1352 (2000).
104. Nishimura, T. et al. CRMP-2 regulates polarized
Numb-mediated endocytosis for axon growth. Nature
Cell Biol. 5, 819–826 (2003).
105. Kawano, Y. et al. CRMP-2 is involved in kinesin-1dependent transport of the Sra-1/WAVE1 complex and
axon formation. Mol. Cell Biol. 25, 9920–9935 (2005).
106. Kobayashi, K. et al. p140Sra-1 (specifically Rac1associated protein) is a novel specific target for Rac1
small GTPase. J. Biol. Chem. 273, 291–295 (1998).
107. Schenck, A. et al. CYFIP/Sra-1 controls neuronal
connectivity in Drosophila and links the Rac1 GTPase
pathway to the fragile X protein. Neuron 38,
887–898 (2003).
108. Miki, H., Suetsugu, S. & Takenawa, T. WAVE, a novel
WASP-family protein involved in actin reorganization
induced by Rac. EMBO J. 17, 6932–6941 (1998).
109. Cole, A. R. et al. GSK-3 phosphorylation of the
Alzheimer epitope within collapsin response mediator
proteins regulates axon elongation in primary
neurons. J. Biol. Chem. 279, 50176–50180 (2004).
110. Uchida, Y. et al. Semaphorin3A signalling is mediated
via sequential Cdk5 and GSK3β phosphorylation of
CRMP2: implication of common phosphorylating
mechanism underlying axon guidance and Alzheimer’s
disease. Genes Cells 10, 165–179 (2005).
111. Arimura, N. et al. Phosphorylation by Rho kinase
regulates CRMP-2 activity in growth cones. Mol. Cell.
Biol. 25, 9973–9984 (2005).
112. Hanger, D. P., Hughes, K., Woodgett, J. R., Brion, J. P.
& Anderton, B. H. Glycogen synthase kinase-3 induces
Alzheimer’s disease-like phosphorylation of tau:
generation of paired helical filament epitopes and
neuronal localisation of the kinase. Neurosci. Lett.
147, 58–62 (1992).
113. Mandelkow, E. M. et al. Glycogen synthase kinase-3
and the Alzheimer-like state of microtubule-associated
protein tau. FEBS Lett. 314, 315–321 (1992).
114. Lucas, F. R., Goold, R. G., Gordon-Weeks, P. R. &
Salinas, P. C. Inhibition of GSK-3beta leading to the
loss of phosphorylated MAP-1B is an early event in
axonal remodelling induced by WNT-7a or lithium.
J. Cell Sci. 111, 1351–1361 (1998).
115. Zumbrunn, J., Kinoshita, K., Hyman, A. A. &
Nathke, I. S. Binding of the adenomatous polyposis
coli protein to microtubules increases microtubule
stability and is regulated by GSK3 β phosphorylation.
Curr. Biol. 11, 44–49 (2001).
116. Mandelkow, E. M. et al. Tau domains, phosphorylation,
and interactions with microtubules. Neurobiol. Aging
16, 355–362; discussion 362–353 (1995).
117. Zhou, F. Q., Zhou, J., Dedhar, S., Wu, Y. H. &
Snider, W. D. NGF-induced axon growth is mediated
by localized inactivation of GSK-3β and functions of
the microtubule plus end binding protein APC. Neuron
42, 897–912 (2004).
118. Takei, Y., Teng, J., Harada, A. & Hirokawa, N. Defects
in axonal elongation and neuronal migration in mice
with disrupted tau and map1b genes. J. Cell Biol. 150,
989–1000 (2000).
119. Baas, P. W. Microtubules and neuronal polarity:
lessons from mitosis. Neuron 22, 23–31 (1999).
120. Kim, W. Y. et al. Essential roles for GSK-3 in
neurotrophin-induced and hippocampal axon growth.
Neuron 52, 981–996 (2006).
121. Al-Khouri, A. M., Ma, Y., Togo, S. H., Williams, S. &
Mustelin, T. Cooperative phosphorylation of the tumor
suppressor phosphatase and tensin homologue
(PTEN) by casein kinases and glycogen synthase
kinase 3β. J. Biol. Chem. 280, 35195–35202 (2005).
122. Zmuda, J. F. & Rivas, R. J. The Golgi apparatus and
the centrosome are localized to the sites of newly
emerging axons in cerebellar granule neurons in vitro.
Cell Motil. Cytoskeleton 41, 18–38 (1998).
123. Lefcort, F. & Bentley, D. Organization of cytoskeletal
elements and organelles preceding growth cone
emergence from an identified neuron in situ. J. Cell
Biol. 108, 1737–1749 (1989).
124. de Anda, F. C. et al. Centrosome localization determines
neuronal polarity. Nature 436, 704–708 (2005).
125. Zolessi, F. R., Poggi, L., Wilkinson, C. J., Chien, C. B. &
Harris, W. A. Polarization and orientation of retinal
ganglion cells in vivo. Neural Develop. 1, 1–21
(2006).
An important paper showing the morphological and
molecular differences between retinal ganglion
cells in vivo and in vitro.
NATURE REVIEWS | NEUROSCIENCE
126. Hirokawa, N. & Takemura, R. Molecular motors and
mechanisms of directional transport in neurons.
Nature Rev. Neurosci. 6, 201–214 (2005).
127. Brady, S. T. A novel brain ATPase with properties
expected for the fast axonal transport motor. Nature
317, 73–75 (1985).
128. Vale, R. D., Reese, T. S. & Sheetz, M. P. Identification
of a novel force-generating protein, kinesin, involved
in microtubule-based motility. Cell 42, 39–50 (1985).
129. Kuznetsov, S. A. et al. The quaternary structure of
bovine brain kinesin. EMBO J. 7, 353–356 (1988).
130. Bloom, G. S., Wagner, M. C., Pfister, K. K. &
Brady, S. T. Native structure and physical properties of
bovine brain kinesin and identification of the ATPbinding subunit polypeptide. Biochemistry 27,
3409–3416 (1988).
131. Nakata, T. & Hirokawa, N. Microtubules provide
directional cues for polarized axonal transport through
interaction with kinesin motor head. J. Cell Biol. 162,
1045–1055 (2003).
132. Kimura, T. et al. Tubulin and CRMP-2 complex is
transported via Kinesin-1. J. Neurochem. 93,
1371–1382 (2005).
133. Ferreira, A., Niclas, J., Vale, R. D., Banker, G. &
Kosik, K. S. Suppression of kinesin expression in
cultured hippocampal neurons using antisense
oligonucleotides. J. Cell Biol. 117, 595–606 (1992).
134. Kamal, A., Stokin, G. B., Yang, Z., Xia, C. H. &
Goldstein, L. S. Axonal transport of amyloid precursor
protein is mediated by direct binding to the kinesin
light chain subunit of kinesin-1. Neuron 28, 449–459
(2000).
135. Verhey, K. J. et al. Cargo of kinesin identified as JIP
scaffolding proteins and associated signaling
molecules. J. Cell Biol. 152, 959–970 (2001).
136. Tsuboi, D., Hikita, T., Qadota, H., Amano, M. &
Kaibuchi, K. Regulatory machinery of UNC-33
Ce-CRMP localization in neurites during neuronal
development in Caenorhabditis elegans.
J. Neurochem. 95, 1629–1641 (2005).
137. Morfini, G., Szebenyi, G., Elluru, R., Ratner, N. &
Brady, S. T. Glycogen synthase kinase 3
phosphorylates kinesin light chains and negatively
regulates kinesin-based motility. EMBO J. 21,
281–293 (2002).
138. Toriyama, M. et al. Shootin1: a protein involved in the
organization of an asymmetric signal for neuronal
polarization. J. Cell Biol. 175, 147–157 (2006).
139. Solecki, D. J., Govek, E. E., Tomoda, T. & Hatten, M. E.
Neuronal polarity in CNS development. Genes Dev.
20, 2639–2647 (2006).
Acknowledgements
We thank S. Kuroda (University of Tokyo) for critical reading
and helpful discussion on the manuscript, T. Miyata (Nagoya
University) for the helpful discussion, and L. Van Aelst (Cold
Spring Harbour Laboratory) for sending a preprint of work in
the press. We also thank T. Yoshimura, A. Hattori, C. Lee,
A. Shimada, N. Mishima and the members of our laboratory
for preparing the manuscript.
Competing interests statement
The authors declare no competing financial interests.
DATABASES
The following terms in this article are linked online to:
Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.
fcgi?db=gene
aPKC | CRMP2 | DOCK7 | FZ | GSK3β | N-WASP | PAR3 | PAR6 |
RAP1B | STEF | TIAM1 | UNC-6 | WNT
FURTHER INFORMATION
Kaibuchi’s laboratory: http://www.med.nagoya-u.ac.jp/
Yakuri/yakuri_en_index.html
SUPPLEMENTARY INFORMATION
See online article: S1 (movie)
Access to this links box is available online.
VOLUME 8 | MARCH 2007 | 205
© 2007 Nature Publishing Group